background filament networks (Murray mesh) in deep sky photos-- noise artifacts or early cosmic structure? Boehringer: Murray 2004.06.15 rmforall
From: Rich Murray (rmforall_at_att.net)
Date: 06/17/04
- Next message: Jon Kickerston: "Re: Kerry on Bush Space Plan: At least he's in favor of Saving Hubble!"
- Previous message: Greg Crinklaw: "Re: Kerry on Bush Space Plan: At least he's in favor of Saving Hubble!"
- Messages sorted by: [ date ] [ thread ]
Date: Wed, 16 Jun 2004 21:35:50 -0600
http://groups.yahoo.com/group/AstroDeep/6
background filament networks (Murray mesh) in deep sky photos-- noise
artifacts or early cosmic structure? Boehringer: Murray 2004.06.15 rmforall
2004 June 15
Bob, thanks for your lucid and careful comments about the
very reasonable interpretation that the faint background filaments noticable
in the background of many extremely deep space photos at very high red
shifts are possibly just noise.
In response to your second post on the effects of raising gamma on making
random noise more visible in photos, I want to point out that the "Murray
mesh" threads are visible on deep sky astronomical photos with gamma at the
usual value 1.00, as in the case of the recent Hubble images of Abell 1689.
Why would random pixels become a mesh of thin, long, crooked, continuous
threads?
With a low-cost program MGI PhotoSuite 4, it is easy to use my 1.4 GH
Pentium 4 system to switch from gamma .30 to 3.00 in steps of .10 gamma, and
magnification from .25 to 4.00, examining the variously colored more
prominent threads. They are quite persistent.
They look to me like a scale invariant, fractile mesh, as predicted by
current models of the evolution of initial large scale structure into
filament networks around voids.
If this is what we are viewing, then of course we are looking from inside a
dense large-scale condensed region, within which our local cluster of
galaxies has evolved, and so it would be expected that we would see the same
dense network of filaments in all directions at a redshift earlier than the
condensation of stars and galaxies in our region, since it is improbable
that we would be located near the boundary of our region. So, the question
of whether "Murray mesh" is artifactual or factual is worth exploring.
The Touchup Filters also includes Invert, which reverses all the colors.
Switching every second to Invert and back makes it easy to find delicate
threads that are visible in both modes.
Would you create a completely random 2 MB image with pixels evenly shared
among white, black, violet, blue, green, yellow, orange, red, and put it on
a site where anyone can copy it and look for similar artifacts?
The striking 3D effect of looking with relaxed, slightly crossed eyes for a
while at paired identical images, until a third image emerges between them,
is my direct experience. I have successfully guided many others to also
have it. It is quite striking to pick out two color postcards in a shop and
see the third 3D image hanging in space between them. I surmise that some
level of the brain's image processing is cued by the slight convergence of
the eyes to carry out the 3D interpretation function, even though the images
are identical.
I find that I see much more in astronomical photos this way, even though the
3D quality may be somewhat off-- for instance, craters may be confused with
hills. I'm interested in whether you and others report success or not in
actually trying it with a number of photos. When I rotate both images 90
degrees or 180 degrees, I get the same effect-- a definite and enjoyable
enhanced perception of the third image, quite different from focusing both
eyes on a single image.
I welcome civil debate on these images.
Anyone can post to AstroDeep@yahoogroups.com
I had to place low-resolution JPG images in the archive at my discussion
group http://groups.yahoo.com/group/AstroDeep/1
However, the filaments are even more convincingly obvious in the original
TIFF images, up to 100 MB of color coded data, especially if the gamma is
shifted from 1.00 to 2.00 or 3.00.
They show up in images of different sky locations, with different
wavelengths, various color codings, and from a variety of large telescopes.
They are easy to see in color deep sky photos in popular astronomy
magazines, especially when the gamma has been shifted to about 2.00 to
render the black background more luminous.
Try it with selections from the recent Hubble photo of Abell 1689 at
http://hubblesite.org/newscenter/newsdesk/archive/releases/2003/01/image/
http://hubblesite.org/newscenter/newsdesk/archive/releases/2003/01/image/f
This 1.89 MB TIFF image at 1X or 2X clearly show myriad red and black
filaments in what I intrepret as a deep 3D mesh, which can been seen behind
as well as between the translucent foreground galaxies. I haven't seen any
signs of gravitational lensing, which might be because they are too far
behind the lensing cluster to be "in focus".
The 3D effect with paired identical color photos is very striking for me-- I
can do it with tourist picture postcards in a store. I'm interested in how
many people can readily experience various versions of 3D perception this
way.
In mutual service, Rich
****************************************************************
From: "Rich Murray" <rmforall@att.net>
Subject: Hubble sees via Abell 1689 to 2 B ly 1.7.3
Date: Tuesday, January 07, 2003 5:01 PM
Hubble sees via Abell 1689 to 2 B ly 1.7.3
Hubble Sees Deep Universe Using Cosmic 'Zoom Lens'
Updated 2:32 PM ET January 7, 2003
By Deborah Zabarenko
SEATTLE (Reuters) - Using a cosmic "zoom lens" made up of cluster of a
trillion stars, the Hubble Space Telescope looked back in time to see
the universe just 2 billion years after the theoretical Big Bang,
astronomers said on Tuesday.
Hubble's new Advanced Camera for Surveys looked straight through a
massive galaxy cluster known as Abell 1689. The gravity of the cluster's
trillion stars acts as a monster magnifying glass in space, warping and
magnifying the light of galaxies far behind it.
Abell 1689 is 2.2 billion light-years away, and it acts as a 2
million-light-year-wide "zoom lens" in space, scientists said at a
meeting of the American Astronomical Society meeting in Seattle.
A light-year is about 6 trillion miles, the distance light travels in a
year.
The new image appears at first glance as hundreds of jewel-like bright
objects -- distant galaxies -- against a black background, much like
previous Hubble pictures.
On closer examination, there are faint arcs of red and blue, the light
from even more remote galaxies smeared by the gravitational bending of
the light as it is magnified.
"We create a kind of pothole in the geometry of the universe," said
Narciso Benitez of the Johns Hopkins University, referring to the
warping known as gravitational lensing.
Some of these galaxies have been seen before, but the new picture
reveals 10 times more arcs than would be seen by a telescope on the
ground, and makes an image twice as sharp as previous images from the
orbiting Hubble's earlier cameras.
Hubble scientists also showed a new image of the dusty disk around a
nearby baby star where planets could lurk.
The 5 million-year-old star -- a true infant in cosmic terms -- lies 320
light-years away in the constellation Libra and appears to be part of a
triple-star system.
Earlier Hubble images showed two rings separated by a dark lane in the
star's disk, and this was interpreted as evidence of one or more planets
around the star.
The new disk image gives a more complex picture, revealing a tight
spiral structure with two arms, one of which appears to be associated
with a nearby double star system.
In a color image, there is a black blob where the light from the star
has been masked to highlight the disk.
"In the picture, we're seeing an interaction between the binary system
and the disk," said Holland Ford, also of Johns Hopkins. "We're not
seeing planets in this disk, but there is nothing that would preclude
planets in this debris disk."
Hubble images and information are available at
http://hubblesite.org/news/2003/01
http://hubblesite.org/newscenter/newsdesk/archive/releases/2003/01/image/
****************************************************************
From: "Louise and Bob" <coatsbob@yahoo.com>
To: "Rich Murray" <rmforall@comcast.net>
Subject: Re: a friendly introduction: Beohringer: Murray 2004.06.12
Date: Monday, June 14, 2004 10:11 PM
I went to the Yahoo groups and took a look at the
photos. I have to say that I did not see any
filaments. I was expecting to see something a little
more obvious.
Having worked for several years in image processing I
was wondering how much of the photos you started with
were information and how much was noise.
For example, if an image has 8 bits of depth per pixel
are there 4, 5, 6, or 7 bits f information. It is
rare to have the lowest bit as actual information and
in many imaging situations there are fewer than 6 bits
of actual information. This means that the lowest 1,
2, and often 3 bits are noise, ie not information.
Image processing is often the task of making images
more pleasing to the eye. This can be simple as in
changing the brightness or contrast. Gamma is a
simple change of pixel intensity in which the change
is greater to darker pixels than to brighter pixels.
Suppose that the images are 6 bit images. That leaves
2 bits of noise. Alter the images so that 2 bits are
information and 2 bits are noise and 4 bits are now
nothing. What just happened to the information
content of the image relative to the noise content?
Also, these images are JPGs. That is a lossy
compression method. How has that changed the noise
content of the images?
Check with the original source of the images to learn
how much of the images is real and how much is noise.
Knowing the quality of the images is important.
Calibrating digital equipment is tricky. Lots of
different techniques have been employed to adjust
sensing equipment.
I am a bit curious about the means of viewing 3-d when
the images are identical such as in viewing multiple
tv screens. The composition of images into an
internal 3-d view by the brain requires that the eyes
see slightly different images. I sometimes work in a
virtual reality lab here at VT. The CAVE produces 3-d
worlds by supplying 24 image pairs to the eyes per
second. A different image is rendered for each eye.
The same is true when a head mounted dsplay is used.
Three-d movies do the same. Two televisions side by
side or two telescope images side by side are not
going to create the 3d effect sicne the images do not
differ.
SIRDS (single image random dot stereograms) produce a
3-d effect, but rely on the use of noise and low
resolution images to produce the effect. Two
superimposed images are laid over a noisy background.
The eyes separate the low resolution images from the
noisy background. The important point here is that
two differing images are imbedded inthe single image.
This is still different than a tv image.
************************************************
From: "Louise and Bob" <coatsbob@yahoo.com>
To: "Rich Murray" <rmforall@comcast.net>
Subject: Re: gamma
Date: Tuesday, June 15, 2004 8:43 AM
The important first step is to consider whether or not
the pattern in the images is real or not. This is
before there is any discussion of redshift or distance
or UV or anything else.
So now I am going to avoid discussing tangled webs of
distant objects and gravitational lensing
possibilities and everything else like that.
Step 1 is to see what gamma is all about. A common way
of computing gamma is as follows:
new = ((old/max)^gamma)*max
Here max is the maximum value of a pixel. For
simplicity of discussion consider max to be 255 which
corresponds to the largest value when 8 bits are used.
Dividing a pixel by max maps the pixels from 0 to max
to the interval 0 to 1. Then the value is raised to
the gamma. The number is still between 0 and 1.
Multiplying the result by max stretches the data back
out to the 0 to max range, which here is 255. So back
to the original range of a pixel.
The gamma value is not the number you entered.
Typically it is 1/n, where n is the number you
entered. Because gamma correction is a point process
it is possible to precalculate what a pixel maps to.
By point process it is meant that each pixel is
independent of its neighbors. Local values do not
affect the result. Each pixel is on its own. I
attached graphs of the correction for gamma values of
2.5, 100, 200, and 300.
Here is what happens. The 2.5 graph is a typical
upper limit for corrections used with monitors. The
other graphs are effectively identical. Take a look
at the left side in the range 0 to 8. Assume 3 bits
of noise. Noise is now raised to the level of bright.
The 5 bits above the noise are all 0. There should
not be anything in the image, yet applying gamma
effectively shoves the noise pixels into the visible
range.
I think it is rather clear that the use of large gamma
values on these pictures is generating nothing, but
noisy images.
If you want to dispute this you might try something
like the following:
Take all of the pixels in the image. Count how many
bits are on and off. For example, take the highest
bit. How often is it 1 and how often is it 0? If the
bit is a noise bit it is conceivable that the number
times that the bit is 1 is approximately equal to the
number of times it is 0. Compare these results to the
results for lesser bits.
There is no 3-d possibility from identical images.
The 3-d effect is based on the differences, apparent
shifts, between the objects in the images. There are
lots of optical tricks that can be played on the mind.
Try this one. Take a pattern of random dots.
Duplicate it. Shift the second patterns a small
amount relative to the first random dot pattern. What
do you see? Is this real or an artifact of the way in
which the visual system tries to make sense of
potential patterns.
bob
****************************************************************
a friendly introduction: Beohringer: Murray 2004.06.12
2004 June 12 Hello Bob,
I enjoyed your website on stereology. You might be interested in my simple
analysis of mysterious background filaments in very deep cosmological
photos, for which I set up a group over two years ago. I am organizing
myself to offer a post on recent images.
"About two
decades ago I noticed that when the same photo is set up side
by side, and viewed with slightly crossed eyes to make a third
composite image in between, that image is created by the
brain's visual system as an excellent 3D image. In fact, you
can visit a TV store, where a lot of sets are all on the same
channel at once, and find two sets the same size, side by side,
and watch the composite image in moving 3D. If you settle
your gaze gently for a few minutes into the composite
image, the innate image processing facility of the brain's
visual system will develop and deepen the 3D appreciation
in remarkable and beautiful ways."
deep sky background filaments: images and interpretation 2002.01.19:
Murray rmforall
http://groups.yahoo.com/group/AstroDeep/1
http://photos.groups.yahoo.com/group/astrodeep/lst?.dir=/&.view=t
Click on the thumbnail photos to get the photos, and click on those
in turn to get full screen photos.
Artifacts? Or?-- immense filaments of H, He, and dark
matter, lit by intense UV from the earliest very massive
stars, "...during the first 10E8 years of the history of
the universe at redshifts between 50 and 10...,"
Prof. Richard B. Larson, Sci. Am. Dec 2001, and
http://www.astro.yale.edu/larson/papers/Noordwijk99.pdf
[7 pages]. This very early intense UV is now redshifted
into the visible and IR bands, and may supply about half of
the current cosmic IR background. The filaments are
generally as thin as 1 pixel.
Photo #2: deeptt1k.jpg:
One pixel = .258 arc-sec, about .25 mm on my 15" monitor.
In MGI PhotoSuite 4.0, I can zoom in to 1600 %, at which point
each pixel is about 4 mm on my 15" monitor.
This is a 20KB cut from the center of the
673 KB original, Photo #1: deeptt1.jpg:
1024X1024 pixels, a random sample, the first of three,
a little to the lower left of center of the 1.15X1.15 degree field,
16000X16000 pixels, 750.3 Mb 24-bit color TIFF,
the highest available resolution,
http://www.noao.edu/image_gallery/html/im0637.html
National Optical Astronomy Observatory Deep Wide-Field Survery.
****************************************************************
----- Original Message -----
From: "bob" <rboehrin@vt.edu>
Newsgroups: bionet.neuroscience
Sent: Tuesday, May 25, 2004 5:43 AM
Subject: Use of stereology
> How often do people make use of stereology in their research. If you
> do use it, do you use a software package or do you use a manual
> technique?
http://filebox.vt.edu/users/rboehrin/index.htm
http://filebox.vt.edu/users/rboehrin/Introduction/AboutAuthor.htm
The Author
I am Robert Boehringer and presently living in Blacksburg, Virginia where I
attend Virginia Tech. I have been enrolled in the Masters program for
Computer Science as a part time student. My GPA is 3.90 (A=4.0 A-=3.7)
Although part time I am an active student and attend as many of the lecture
series as possible. I also make good use of the cultural opportunities that
are available through the university.
I am employed by MicrobrightField, Inc. the leader in software for
stereological research and serial reconstructions.
Outside of the school and work I have hobbies that include bird watching,
hiking, traveling, and the occasional rock climb.
My motivation for creating information about stereology is due to the the
lack of information available online. It also provides me the chance to
record some of the observations I have made.
It is surprisingly easy to find misinformation about stereology. Examples
are:
Suggestions to avoid proper sampling
Poorly done simulations
The use of ocular mathematics
The latter entry is in reference to a joke that several of us started in
high school. It was suggested that the easiest math would be a discipline
with a single axiom, "If it looks right, then it is right." It should come
as no surprise that ocular mathematics is prevalent in many disciplines.
Stereology is no exception. I have seen corrections for numbers that are too
large, generate even larger numbers. I have seen counting rules changed to
forms that were more pleasing to the eye. I have seen assumptions made about
averages that do not hold under even the simplest conditions. I cannot be
certain in all cases, but I believe that the ocular axiom was invoked in all
of these cases as well as many others.
Please send in suggestions or comments to rboehrin@vt.edu.
Tricouni Nail in the Needles of South Dakota
This page written by Robert Boehringer at Virginia Tech.
****************************************************************
http://groups.yahoo.com/group/aspartameNM/message/1071
research on aspartame (methanol, formaldehyde, formic acid) toxicity:
Murray 2004.06.15 rmforall
Rich Murray, MA Room For All rmforall@comcast.net
1943 Otowi Road, Santa Fe, New Mexico 87505 USA 505-501-2298
[ NutraSweet, Equal, Canderel, Benevia, E951 ]
http://groups.yahoo.com/group/aspartameNM/message/927
Donald Rumsfeld, 1977 head of Searle Corp., got aspartame FDA approval:
Turner: Murray 2002.12.23 rmforall
****************************************************************
- Next message: Jon Kickerston: "Re: Kerry on Bush Space Plan: At least he's in favor of Saving Hubble!"
- Previous message: Greg Crinklaw: "Re: Kerry on Bush Space Plan: At least he's in favor of Saving Hubble!"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|
